The observation of atomic-scale structures of grain boundaries (GBs) via axial illumination HREM has been largely restricted to tilt GBs, due to the requirement that the electron beam be parallel to a low-index zone axis on both sides of the interface. This condition can be fulfilled for all tilt GBs with misorientation about a low-index direction. The information obtained through HREM studies in many materials has brought important insights concerning the atomic-scale structure of such boundaries. However, it is well known that tilt GBs occupy only an infinitesimally small fraction of the 5-dimensional phase space which describes the macroscopic geometry of all GBs. Therefore, although tilt GBs are very important due to their low energy, it would be usefulto also study twist GBs and general GBs that contain twist and tilt components.

We have prepared thin-film Au samples by an epitaxy technique in which (01l) and (001) grains are grown side by side.

The electrical activity of grain boundaries in perovskites, of which SrTiO3 is a model system, is the basis for their use as capacitors, varistors and positive-temperature coefficient resistors. In related materials this same electrical activity is often detrimental. The outstanding example of a negative effect is the reduction in critical current by several orders of magnitude across grain boundaries in YBa2CU3O7-δ as the boundary misorientation is increased from 0°-45°. Grain boundaries are also likely to profoundly affect properties such as domain wall motion in ferroelectric and magnetic perovskites. The origin of the electrical activity is ubiquitously attributed to the existence of grain boundary donors, usually assumed to be impurities, which set up a double Schottky barrier as they are screened by dopants in the adjacent bulk crystal. We show here, that although electrical barriers can doubtless be modified by dopants, the grain boundary structure itself is the intrinsic origin of the socalled grain boundary donors.

Ferroelectric thin films have received much attention because of their nonvolatile ferroelectric random access memory (NVFRAM) applications. Recently SrBi2Ta2O9 (SBT), a bismuth-layered ferroelectric compound, has been recognized as a leading candidate for the NVFRAM applications because of its negligible fatigue, low leakage currents and ability to maintain bulk characteristics when fabricated in very thin films. While most work up to date has been focused on its thin film forms for nonvolatile memory applications, its intrinsic properties, basic structure and the effects of defects on the properties are not well understood, mainly due to the lack of sizable single crystals and the details of defects. In this paper we report the first direct determination of a novel planar defect structure in SrBi2Ta2O9, using high-resolution Z-contrast imaging.

The n+1TinO3n+1 Ruddlesden-Popper homologous series are promising candidates in the search for low loss tunable dielectric materials. SrTiO3 (n=∞) has the perovskite structure with alternating TiO2 and SrO layers. However, those members of the n+1TinO3n+1 series with n≠∞ have tetragonal symmetry and space group 14/mmm. These phases are formed by inserting a rock-salt SrO layer every n SrTiO3 perovskites, resulting in a successive perovskite blocks being mutually sheared by [1/2,1/2,0] (Fig. 1).

Previously, the n = 1 - 3 members of this series have been made in polycrystalline form using solid state synthesis methods. Attempts to synthesize intermediate members (4≤n<∞) have been unsuccessful and resulted in mixed phase samples. Fortunately, significant advancement in epitaxial growth techniques makes it possible to control the synthesis of a wide variety of homologous oxide crystal systems. In the present work, thin films of the Srn+1TinO3n+1, RP series with n = 1 - 5 have been synthesized using reactive molecular beam epitaxy (MBE).

A vanadium-bearing variant of 316 stainless steel that was rapid solidification processed (RSP) by gas atomization and hot extrusion of the powder (10:1 extrusion ratio at 900°C) has been studied previously by conventional atom probe field ion microscopy (APFIM). The mechanical properties of this steel were markedly improved by RSP and aging (600°C for 1000 hours). High nitrogen (0.45 at% (0.2 wt%)) and oxygen (0.16 at% (0.05 wt%)) contents were intentionally introduced by melting under 80% nitrogen/20% oxygen gas and atomizing in nitrogen gas. A nominal boron concentration of 0.04 at% (0.01 wt%)) is present as a tramp element. As a result, a large number density (˜ 2 × 1021 m-3) of 25 nm plate-like vanadium-rich nitrides precipitate during aging of the alloy and these precipitates contribute a major portion of the strengthening. Previous efforts to locate the oxygen in the structure using APFIM were inconclusive largely due to poor counting statistics.

The technique of atom probe tomography (APT) enables the x, y, and z coordinates and the elemental identities of the atoms in a small volume to be determined at the atomic level. Therefore, the APT technique may be used to characterize solute segregation to interfaces and precipitation in terms of concentration gradients and precipitate morphology. This type of information may be used to optimize the design of alloys.

The material that was used to illustrate the capabilities of atom probe tomography is a complex polycrystalline nickel-based superalloy, Alloy 718. The composition of this commercial superalloy is Ni- 3.2 at. % Nb, 0.96% Al, 1.15% Ti, 20.3% Fe, 21.8% Cr, 0.26% Co, 1.8% Mo, 0.16% Mn, 0.21% Si and 0.26% C. The material was characterized after a heat treatment oM h at 1038°C + 8 h at 870°C + 500 h at 600°C. Previous atom probe field ion microscopy characterizations of this material has demonstrated that there is no intragranular precipitation after the anneal at 1038°C.

The transistors planned for commercial use ten years from now in many electronic devices will have gate lengths shorter than 130 atoms, gate oxides thinner than 1.2 nm of SiO2 and clock speeds in excess of 10 GHz. It is now technologically possible to produce such transistors with gate oxides only 5 silicon atoms thick[l]. Since at least two of those 5 atoms are not in a local environment similar to either bulk Si or bulk SiO2, the properties of the interface are responsible for a significant fraction of the “bulk” properties of the gate oxide. However the physical (and especially their electrical) properties of the interfacial atoms are very different from .bulk Si or bulk SiO2. Further, roughness on an atomic scale can alter the leakage current by orders of magnitude.

In our studies of such devices, we found that thermal oxidation tends to produce Si/SiO2 interfaces with 0.1-0.2 nm rms roughness.

A number of recent studies of grain boundaries and heterophase interfaces have demonstrated the power of combining Z-contrast STEM imaging, EELS and first-principles theoretical modeling to give an essentially complete atomic scale description of structure, bonding and energetics. Impurity sites and valence can be determined experimentally and configurations determined through calculations.

Here we present an investigation of the Si/SiO2 interface. The Z-contrast image in Fig. la, taken with the VG Microscopes HB603U STEM, shows that the atomic structure of Si is maintained up to the last layers visible. The decrease in intensity near the interface could originate from interfacial roughness of around one unit cell (∼0.5 nm), or may represent dechanneling in the slightly buckled columns induced by the oxide. Fig. lb, taken from a sample with ∼1 nm interface roughness, shows a band of bright contrast near the interface. This is not due to impurities or thickness variation since it disappears on increasing the detector angle from 25 mrad to 45 mrad (Fig. lc), and is therefore due to induced strain.

SIMS has inherent difficulties with quantification because of the so called “matrix effect”. Many factors contribute to the matrix effect, e.g. differing concentration of oxygen, sputtering rate differences in the hetero-layers, etc. In the case of MOS(metal-oxide-semiconductor) structures, the oxide layer gives rise to a large matrix effect. It is thus very difficult to use SIMS to evaluate the relationship between the electrical properties of the LSI devices and the impurity profiles present in such systems.

So we have been studying laser post-ionization SNMS, which consists of TOF-SIMS apparatus and excimer laser, in order to quantify the impurity profiles around SiO2/Si interface. Depth profiles of implanted Cu with 1×1015 atoms/cm2 in SiO2(100 nm)/Si system taken by Monte-Carlo simulation ,SIMS and laser post-ionization SNMS are shown in Fig.l. In this implantation condition the Cu+:Cu2+:Cu3+ percentage ratios of the charge distributions were 44:42:14. Fig. 1(a) shows the theoretical depth profiles expetted from this implantation condition by Monte-Cairo simulation.

Growth of germanium or silicon-germanium on silicon 100 surfaces produces pyramid shaped islands after several monolayers of coverage. With further deposition, these islands will continue to grow until they reach a critical volume where a transition to a dome shape occurs. The critical surface coverage for island formation and the critical volume for the pyramid to dome transition depend upon the silicon to germanium ratio or degree of mismatch between the deposited material and the silicon substrate. Strain plays a dominant role in the shape transitions. Annealing studies have shown that the pyramid to dome transition.is reversible and is likely due to strain reduction related to interdiffusion. Annealing results in island growth, a decrease in the number of islands, and interdiffusion of the silicon and germanium. The interdiffusion relieves the strain and shifts the critical volume to a value exceeding the current volume of the dome shaped island thereby driving the transition back to pyramidal shape.

Changes in electronic structure and bonding at metal-oxide interfaces can have a pronounced influence on the macroscopic properties of the interface. This is particularly true for supported-metal catalysts, which feature nanometer-sized transition metal particles dispersed on an inert support such as A12O3 or SiO2. Variations in the d-band filling of the transition metal due to chemical bonding with the support could potentially alter the chemical activity of metal atoms near the interface, thereby affecting catalyst performance. This work investigates possible interface contributions to the performance of a model Ni- SiO2 catalyst system by directly assessing the nature of bonding at the metal-oxide interface. Electron energy-loss spectrometry (EELS) in the TEM is used to examine and quantify changes in the d-band occupancy of interfacial Ni atoms relative to the bulk, and to further relate these changes to interface chemistry.

The Ni d-band occupancy is evaluated using the integrated area under the Ni L2,3 edge in the energyloss spectrum, according to routines establishe-i previously.

GaN is a direct, wide band -gap semiconductor (Eg =3.4 eV) which has a potential application in electronic and optical devices including UV-emitting lasers and bright p-n junction GaN light emitting diodes. In order to attain optimum device performance, it is necessary to develop lowresistance, thermally stable and uniform ohmic contacts.

In this study, the TiN/GaN contact synthesis and examination of contact performance were conducted at North Carolina State University. Si-doped n-GaN films were grown on 6H-SiC substrates via MOVPE with an A1N buffer layer. The TiN films were deposited in an ion-beam assisted, UHV electron-beam evaporation system. TiN growth was performed at a substrate temperature of 350 °C and a nitrogen pressure of about 2×10-4 torr. The deposition rate was 1- 1.5 nm/min. The TiN contacts deposited on (0001) n-GaN were ohmic with high resistivity in the as-deposited condition, but the resistivity decreased sharply in response to annealing. The TiN/GaN interface was examined by transmission electron microscopy to provide correlations between changes in microstructure and electrical behavior.

In this proceedings, we present a review of our experimental results of our investigations of the mechanisms of the initial stages of copper oxidation. We examined the initial stages of Cu(001) oxidation and reduction by in situ ultra-high vacuum (UHV) transmission, electron microscopy (TEM). We observed surface reconstruction and nucleation and growth of copper oxide islands. We have examined the oxidation processes from oxygen partial pressures of 10-5 torr to atmospheric pressures and temperatures from 25°C to 600°C, in order to gain fundamental insights into this important gas-metal reaction.

Fundamental knowledge of gas-metal reactions, in particular oxidation, is important for a wide variety of materials science fields, such as dry corrosion, catalysis, as well as some thin film growth, such as ferroelectrics. However, there is a wide gap between information provided by surface science methods and that provided by bulk oxidation studies. The former have mostly examined the adsorption of up to ˜1ML of oxygen on the metal surface.

Transmission electron microscopy plays a key role in the investigation of precipitates and inclusions because it provides direct information at high spatial resolution, on their shape, orientation relationship, interface structure and composition.

This work has utilized high resolution and in-situ electron microscopy to study the equilibrium shapes and thermal behavior of Pb precipitates in Al. It was found that PB precipitates within Al grains form in parallel-cube orientation relationship, are cuboctahedral in shape, adopt a sequence of magic sizes, and display an increase in aspect ratio with decreasing size. These observations have been explained in terms of interfacial energy and residual strain energy.

When the same Pb precipitates form on grain boundaries, their shapes are significantly different. This is apparent from figure 1 which shows a 90° <110> tilt grain boundary in Al with a small cuboctahedral Pb precipitate in the upper grain and a large complex-shaped Pb particle on the grain boundary.

Development of high-density longitudinal magnetic recording media (used for computer hard disks) with good noise performance and high thermal stability requires optimization of both alloy composition and processing methods. In CoCr(PtTa) thin films, intergranular Cr segregation is responsible for decoupling the magnetic exchange between the small ferromagnetic grains. The corresponding Cr depletion within the grains affects the “bulk” magnetic anisotropy. However, the nanoscale structural and chemical details that are needed for modeling and for guiding material development are not well understood. Energy-filtered transmission electron microscopy (EFTEM) has been used to characterize a series of CoCrTa/Cr, CoCrPt/Cr, and CoCrPtTa/Cr media sputtered under various processing conditions, in order to understand their structure-property-processing relationships.

Processing typical for hard-disk media was used: 30 or 60-nm films of a Cr underlayer followed by a Co alloy were d.c. magnetron sputtered onto a NiP-plated Al substrate pre-heated to 250°C.

The discovery of amorphous, or glassy, metallic alloys in 1959 posed an intellectual challenge. How can one describe the structure of glasses when there is no long-range periodicity? What can the structure tell us about why certain metal alloys form glasses more easily than others? First, some universal characteristics, if any exist, of the structure metallic glasses needed to be found. A convincing model was proposed for the structure of metallic glasses based on Bernal’s dense random packing (DRP) structure. Central to this proposal is the idea that the structure of metallic glasses is that of the random filling of space by non-interacting identical spheres. In this model, strongly directional interatomic bonds do not play an important role in determining the structure of metallic glasses. This model is hpwever in conflict with one proposed by Chen, which correlates increased glass formability with increased chemical interaction between dissimilar atoms.

Carbon nanotubes and carbon onions are two fascinating forms of crystalline nanosize carbons that have received considerable attention in recent years (1,2). The remarkable physical properties of nanotubes make them valuable material for applications ranging from electronic devices to nanoprobes. The dimensions and topology of these structures make them fascinating objects for study. Multiwalled nanotubes and onions are made of concentric graphite layers, with different geometry. Singlewalled nanotubes on the other hand are made of singular graphene cylinders, often self-assembled into larger rope-like structures. Nanotubes can be produced in gram quantities using electric arc or laser ablation. Carbon onions, however, are structures that assemble from carbonaceous soot under intense electron irradiation in an electron microscope. Hence the onions are metastable structures that become unstable and disintegrate in the absence of irradiation.

The multiwalled carbon nanotubes and carbon onions have been shown to encapsulate foreign materials (metals, oxides, carbides), either during growth or by capillarity filling (in the case of nanotubes) (1), or during irradiation of metal containing soot (onions) (3).

Supported Pt-metal catalysts are of immense commercial importance for hydrogenation reactions. The fine chemicals industry relies on special catalysts that selectively hydrogenate specific bonds or chemical functions on molecules. The activity, selectivity, lifetime, etc. of a supported hydrogenation catalyst is wellknown to depend strongly on the chosen support, but systematic studies of the role of the support and of support-metal interactions are lacking in the literature. The purpose of this work is to examine in detail the function of single-wall carbon nanotubes (SWNT) as support for a Pt catalyst, in the context of a selective hydrogenation reaction. The test reaction consists of the hydrogenation of the carbonyl group on 3-methyl-2- butenal, a reaction extensively studied on well-characterized Pt surfaces and also unsupported polycrystalline Pt. This reaction, which follows a parallel/consecutive reaction network (Fig. 1), represents a class of selective hydrogenation reactions of α,β-unsaturated aldehydes that are important in the fine chemicals industry for flavorings, fragrances, and some pharmaceuticals.

Single wall carbon nanotubes (SWNT) were doped with iodine, resulting in charge transfer between the iodine and carbon [1]. It was found that the iodine intercalation is air stable and also reversible by simple heat treatment. This behavior is in contrast to graphite and fullerenes which do not form charge transfer compounds with iodine. Although the iodine-induced charge transfer was not found in any other carbon polymorph solid, it has been found in some low dimensional organic polymers such as polyaniline,[2] forming charged linear-chain polyiodides (I3)- and (I5)-. This suggests that the geometric configuration of the carbon may play a very important role in iodine intercalation. In this study, we use Z-contrast imaging[3] in a VG Microscopes HB603U STEM to directly observe the configuration of iodine within the nanotube bundles. First-principles density-functional calculations are then used to explain the preference for the observed iodine configuration within the nanotubes.

The quantification of grain boundary segregation levels, as measured with X-ray energy dispersive spectroscopy (XEDS) in a scanning transmission electron microscope (STEM), is dependent on the size and shape of the interaction volume. The segregation level T (in atoms/nm2) is related to the intensities of the characteristic peaks in the X-ray spectrum, Is and Im, by

where ρ is the density of the matrix in atoms/nm3, Am and As are the atomic masses of the matrix and segregant respectively and ksm is the usual k-factor. The geometric factor, V/A, is the ratio of the volume of interaction to the area of the grain boundary inside in the interaction volume. Different models have been used to describe the interaction volume and these are illustrated in Fig. 1 and the appropriate expression for V/A is given in each case. In the simplest case, beam broadening is neglected and the interaction volume can be described as a cylinder with diameter equal to the probe size, d.